Discrimination of dicentric chromosome from radiation exposure patient data using a pretrained deep learning model

• Published in: Nuclear engineering and technology Vol. 56; no. 8; pp. 3123 - 3128
• Main Authors: Kwon, Soon Woo, Jang, Won Il, Kim, Mi-Sook, Seong, Ki Moon, Lee, Yang Hee, Yoon, Hyo Jin, Yang, Susan, Lee, Younghyun, Shim, Hyung Jin
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.08.2024; Elsevier; 한국원자력학회
• Subjects: Biological dosimetry; Convolutional neural network; Dicentric chromosome assay; Patient with radiation exposure; VGG19; 원자력공학; Dicentric chromosome assay; VGG19; Biological dosimetry; Convolutional neural network; Patient with radiation exposure
• ISSN: 1738-5733; 2234-358X
• DOI: 10.1016/j.net.2024.03.011

The dicentric chromosome assay is a gold standard method to estimate radiation exposure by calculating the ratio of dicentric chromosomes existing in cells. The objective of this study was to propose an automatic dicentric chromosome discrimination method based on deep convolutional neural networks using radiation exposure patient data. From 45 patients with radiation exposure, conventional Giemsa-stained images of 116,258 normal and 2800 dicentric chromosomes were confirmed. ImageNet was used to pre-train VGG19, which was modified and fine-tuned. The proposed modified VGG19 demonstrated dicentric chromosome discrimination performance, with a true positive rate of 0.927, a true negative rate of 0.997, a positive predictive value of 0.882, a negative predictive value of 0.998, and an area under the receiver operating characteristic curve of 0.997.